Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
DOI: 10.4018/IJSWIS.2017100104
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
Copyright©2017,IGIGlobal.CopyingordistributinginprintorelectronicformswithoutwrittenpermissionofIGIGlobalisprohibited.
The Use of Software Tools in Linked Data Publication and Consumption:A Systematic Literature ReviewArmando Barbosa, Federal University of Alagoas, Maceió, Brazil
Ig Ibert Bittencourt, Federal University of Alagoas, Maceió, Brazil
Sean Wolfgand Matsui Siqueira, Federal University of the State of Rio de Janeiro (UNIRIO), Rio de Janeiro, Brazil
Rafael de Amorim Silva, Federal University of Alagoas, Maceió, Brazil
Ivo Calado, Federal Institute of Alagoas, Rio Largo, Brazil
ABSTRACT
ToreducethecomplexityintrinsictoLDmanipulation,softwaretoolsareusedtopublishorconsumedataassociatedtoLDactivities.However,fewdevelopershaveabroadunderstandingofhowsoftwaretoolsmaybeusedinpublicationorconsumptionofLinkedData.ThegoalofthisworkistoinvestigatetheuseofsoftwaretoolsinLinkedDatapublicationandconsumptionprocesses.Morespecifically,understandinghowthesesoftwaretoolsarerelatedtoprocessofpublicationorconsumptionofLD.Inordertomeettheirgoal,theauthorsconductedaSystematicLiteratureReview(SLR)toidentifythestudiesontheuseofsoftwaretoolsintheseprocesses.TheSLRgathered6473studies,ofwhichonly80studiesremainedforfinalanalysis(1.25%of theoriginalsample).Thehighlightsof thestudyare:(1)initialstepsofthepublicationprocessarefairlysupportedbythesoftwaretools;(2)Non-RDFserializationisfairlysupportedinpublicationandconsumptionsprocessbythesoftwaretools;and(3)therearenon-supportedstepsinconsumptionandpublicationprocessesbythetools.
KeyWoRDSConsumption Process, Linked Data, Ontology, Publication Process, Software Tools
1. INTRoDUCTIoN
Inlastdecade,theevolutionoftheWebhasincreasedtheconsumptionandproductionofcontent.Thisgrowthimpliesinseveralchallengessuchasalongertimerequiredtoperformsearchesorbrowsing,dataconsistencyproblems,andmisinterpretations.Theusersandmachinesperceivethesechallenges,astheyarenecessarytoaccessasetofwebpagestofindrelevantinformation.Inaddition,mostofthedataavailableontheWebareunstructured(images,blogs,webcontent,etc.),whichmeansthatmachinesarenotable toautomaticallyunderstandsuchdatacontent.Therefore, it isessential toprovidealternativesforstructuringdatainamachineunderstandablemanner.
LinkedData(LD)isaneffectiveoptiontostructuresuchdata,presentingasetofbestpracticesforguidingpublishingandconnectingthesestructureddataontheWeb(Hyland&Wood,2011).LDadoptsagenericabstractdatamodelfordescribingresourcescalledResourceDescriptionFramework
68
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
69
(RDF).Thisframeworkisresponsiblefordescribingtermsandtheirrelationships,byusingtriplesofsubjects,predicatesandobjects.Additionally,ontologiesmayprovideenrichmenttoLinkedDatabyaddingsemanticstotheirRDFstatements(Heath&Bizer,2011).Moreover,LDbringssomebenefitswhichmaybeappliedinmultipleresearchfields,suchasdatainteroperabilityandtermsdefinition(Bauer&Kaltenböck,2011).SomeapplicationsofLD includeOpenGovernment (Ding,Lebo,Erickson,DiFranzo,Williams,Li,&Hendler,2011;Shadbolt,O’Hara,Berners-Lee,Gibbins,Glaser,&Hall,2012)(e.g.toimproveoftransparencyandprovenance),Education(Dietze,Yu,Giordano,Kaldoudi,Dovrolis,&Taibi,2012)(e.g. to improveeducationrepositoriesandcontentsharing),andBusiness(Barnes&Martin,2002)(e.g.toimprovebusinessopportunitiesandbusinessvalue).
LD is defined by both publication and consumption processes. Recently, W3C (2014) haspublishedalistof10stepstofulfillLDpublicationrequirements:(1)topreparestakeholders;(2)toselectdatasetforreuse;(3)tomodelthedatatorepresentdataobjectsandtheirrelationships;(4)tospecifyasuitableopendatalicense;(5)todefineURIforlinkeddata;(6)todescribeobjectswithstandardvocabularies;(7)toconvertdatatoalinkeddatarepresentation;(8)toprovidemachineaccesstodata;(9)toannouncetothepublic;and(10)toestablishasocialcontractofalinkeddatapublisher.Ontheotherhand,BauerandKaltenböck(2011)havepublishedalistof7stepsrelatedtoLDconsumptionprocess:(1)tospecifyconcreteusecasesfornewservicesorapplications;(2)toevaluaterelevantdatasourcesanddatasets;(3)tochecklicensesforuseandreuseprovidedbydataowners;(4)tocreateconsumptionpatternstospecifywhatdataarereusedfromacertaindatasource;(5)tomanagealignment,cachingandupdatemechanisms;(6)tocreatemashupsandapplicationstoprovideuser-friendlygraphicaluserinterfaces(GUIs)andpowerfulservicesforendusers;and(7)toestablishsustainablepartnerships.
Both aforementioned processes are essentials to improve data quality, interoperability, anddiscoverabilityofdata, aswell as to support thedevelopmentof richapplications.Additionally,consideringthenumberofstepsandactivities,itisimportantthedevelopmentoftoolsforsupportingLinkedDatapublicationandconsumptionprocesses.ConsideringthehugenumberofLinkedData-relatedtoolsproposedrecently(Bizer,Heath,&Berners-Lee,2009),itisatime-consumingtask,forpublishersorconsumers,toknowtheavailablefeaturesofeachtoolinordertomakeproperlydecisionaboutsoftwaretoolsthatshouldbeusedaccordingtoeachstep.Forthisreason,ahigh-levelquestionwasraised:Whichsoftwaretoolshavebeenusedtosupportthepublicationand/orconsumptionofLinkedData?Inthiscontext,ourpurposeistoidentifywhattoolshavebeenusedtoperformtheaforementionedtasks,providingabroadvisionaboutLinkedDatasoftwaretoolstopublishand/orconsumelinkeddata.Additionally,weusethecategoriesprovidedbyW3C1toclassifythesoftwaretoolsaccordingtotheirfeatures.
TherearesomesimilarworksthatinvestigatedtheuseofsoftwaretoolstopublishorconsumeLinkedData(Corlosquet,Delbru,Clark,Polleres,&Decker,2009;Bizeretal.,2009).However,thesestudiesdonotcaptureall theaspectsandevidencesthatweareinterested,suchassupportandformats.Moreover, theirworkissomewhatlimited,astheydonotsystematicallyreviewtheliteraturetoinvestigatetheworksthatjointlycoversoftwaretoolsandLinkedDataPublicationandConsumptionProcess.
Therefore,inthispaperweusetheSystematicLiteratureReview(SLR)methodtounderstandwhichsoftwaretoolssupportpublicationandconsumptionoflinkeddata.Weutilizethemethoddefinedin(Kitchenham&Charters,2007)toidentify,evaluate,interpretandsynthesizetheavailablestudies in order to answer particular research questions on the symbiosis of software tools forpublicationand/orconsumptionoflinkeddata,andtoestablishthestateofevidencewithanin-depth
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
70
analysis.Thoughttheanalysiswegeneratesixquestions.Thesequestionsmayguidenewresearchline,softwaresolutions,anddiscussion.
Therestofthispaperisorganizedasfollows:Section2describestheSLRmethodusedinthisreviewandprimaryresults.Section3presentsthequalityassessment.Section4presentsresults.Section5presentsadiscussionabouttheresults.Section6presentsthreatstovalidity.Finally,inthelastsectionwepresenttheconclusionsandfutureworkofthisSLR.
2. SySTeMATIC LITeRATURe ReVIeW
Asystematicliteraturereview(SLR)demonstratesthestateoftheartinsomeparticularresearchfield,evaluatingfindingsrelatedtoagivenresearchissue,atopicarea,orevenaphenomenon.ItisworthtodifferentiateaSLRfromaSystematicMapping.AccordingtoKitchenhametal.(2007),SystematicMappingsorScopingStudiesprovideawideoverviewofaresearcharea.TheresultsofamappingstudycanidentifyareassuitabletoconductingaSLRandareaswhereaprimarystudyismoreappropriate.
Inthiswork,weadopttheguidelinesproposedbyKitchenhametal.(2007)toconductourSLR.WealsoadoptthetoolcalledStArt(StateoftheArtthroughSystematicReviews)(LaPES,2012)toconductourSLR,reducingerrorsinherenttotheSLRprocess.Thistoolisvalidatedin(Hernandes,Zamboni,Fabbri,&Thommazo,2012),obtainingahighperformanceontheexecutionofaSLR.
Inthiswork,weinvestigatethefollowingresearchMainQuestion(MQ):
MQ→Whichsoftwaretoolssupportthepublicationand/orconsumptionofLinkedData?
AccordingtoKitchenhametal.(2007),themainstepforanysystematicreviewisthedefinitionoftheresearchquestions,sincetheyareresponsibleforguidingthewholereviewprocess.Basedonsuchassumption,andsimilarlytoothersystematicreviews(Dermeval,Vilela,Bittencourt,Isotani,&Brito,2014),weelaboratethreeresearchquestionsinordertoenableustoanswertheMQ.
• RQ1→Which are the linked data-based publication / consumption steps utilized by software tools in current literature?
• RQ2→Which serialization formats does the tool support?• RQ3→Is the usage of software tool empirically evaluated?
RQ1providesastartingpointtounderstandthesoftwaretoolsandthemainapproach(publish/consume)supportedbythem.Thisquestionisresponsibleforidentifyingthestepsofthepublicationandoftheconsumptionprocesssupportedbythesoftwaretools.RQ2identifiesthemainserializableformats(RDF,OWL,JSON,CSV,XMLandothers)usedforpublishingorconsuminglinkeddata.Finally,RQ3intendstoanalyzeifsuchstudiesprovideanempiricalevaluationofthesoftwaretools.Besides,Table1presentsthecentralmotivationforeachresearchquestion.
ThestepIIdefinesinclusionandexclusioncriteriaadoptedinthiswork.AccordingtoKitchenhametal.(2007),criteriaintheSLRprocessidentifythosepapersthatprovidedirectevidenceaboutresearchquestions,reducingapossiblebias.WeadoptedtwotypesofcriteriaintheSLR:(1)inclusioncriteria;and(2)exclusioncriteria.TheformerdefinesallthenecessaryrequirementsinorderanarticlemustbeconsideredintheSLRprocess.Thelatterdefinesalltherequirementsforanarticlenottobeincludedinthisprocess.Weconsideredprimarypapers(thosethatpresentsomeproposalinthearea)intheinclusioncriteriawhereassecondarypapers(i.e.thosethatonlyreviewanareatopicsuchasaSLR)intheexclusioncriteria.Regardingtopublicationdate,studiesareeligibleforinclusionintheSLRprocesswhethertheywerepeer-reviewedandpublishedfromJanuary2006
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
71
untilDecember2015.Intheotherhand,apaperisexcludedfromtheprocesswhetheritbelongstosomeexclusioncriteriadefinedinTable2.
ToperformthestepIII,weconsideredthefollowingelectronicdatabases:
• ScienceDirect(http://www.sciencedirect.com);• ISIWebofKnowledge(http://apps.webofknowledge.com);• Scopus(htttp://www.scopus.com);• SpringerLink(http://link.springer.com/);• ACMDigitalLibrary(http://dl.acm.org);• IEEEXplore(http://ieeexplore.ieee.org);• EngineVillage(http://www.engineeringvillage.com).
Figure1showsthesystematicreviewprocessandthenumberofpapersidentifiedateachstage.It isworthhighlighting the steps related to thepaper selectionprocess.Thus,at theStep I,
Weextractedthepapersfromelectronicdatabasesusingthefollowingsearchterms:(1)softwaretool terms,containing“Environment”OR“Authoring”OR“Architecture”OR“Framework”OR“Software”OR“Application”OR“Program”;(2)linkeddataterms,containing“LinkedData”OR“LinkedOpenData”;and(3)ontologyterms,containing“Ontology”OR“Ontologies”.Fromthesearchresults(6,473papers),weautomaticallyaccessed,downloadedandorganizedwithaidingoftheStArttool.Thesetermswerecombinedasfollows:
(1)AND(2)AND(3)
Table 1. Research Question Motivation
RQ Motivation Research Question
RQ1Ingeneral,toolssupportonlysomestepsofthepublication/consumptionprocess.Inthiscontext,itisimportanttodetermineforeachanalyzedtooltheexactsetofsupportedsteps.
Whicharethelinkeddata-basedpublication/consumptionstepsutilizedbysoftwaretoolsincurrentliterature?
RQ2 Consideringtherearedifferentserializationformatsavailable,wewanttodeterminetheformatsthateachanalyzedtoolsupports.
Whichserializationformatsdothetoolsupport?
RQ3Theevaluationofapublication/consumptiontoolisessentialtoquantifytheeffectivenessoftheanalyzedtool.So,weintendtoverifywhichempiricalevaluationswereappliedforeachtool.
Istheusageofsoftwaretoolempiricallyevaluated?
Table 2. Inclusion and exclusion criteria
# Inclusion Criteria # Exclusion Criteria
1 Papersabouttoolsforpublishingand/orconsuminglinkeddata 1 PapersinotherlanguagethanEnglish
2 PrimaryStudies 2 ShortPapers
3 ComputerSciencePapers 3 DuplicatedStudies
4 OnlyPapersinEnglish 4 SecondaryStudies
5 PapersbetweenJan.2006andDec.2015 5 Papersonlyabouttools
6 Paperswhichanswertheresearchquestions 6 Papersonlyaboutlinkeddata
7 Papersonlyaboutsoftwaretoolsfocusedopendata(withoutLinkedData)
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
72
AttheStepII,wedetectedandremovedduplicatedpapersusingtheStArttool,remainingasetof5,068papers.
AttheStepIII,wereviewedtitlesandkeywordsofeachpaperandexcludedthosenotrelatedtotheresearchquestions(1,733papers).Theremainingsetofpapers(3,335)wereconsideredforthenextstepsincedatawasinsufficienttodecideonremoval.
AttheStepIV,weanalyzedtheabstractsandexcludedpapersaccordingtotheexclusioncriteria(3,190papers).SimilarlytoStepIII,thepaperswithinsufficientinformationtobeclassifiedontheexclusioncriteriaremainedinthelisttobeanalyzedinthenextStep.
At theStepV,we retrieved thecomplete text from the remainedpapers (144), and readbothintroductionandconclusionsectionsand,additionally,performedafull-textscreenedofeachoneoftheselectedpapers.Fromthisstep,64paperswereexcludedinthisstepaccordingtotheexclusioncriteria.
AfterfinishingStepV,80papersremainedtothedataextractionprocess,whichhastwoactivities:(1)readingthefull-textofeachpaperand(2)savingextracteddata(i.e.inaspreadsheet,database,textfile,etc.).Therefore,allremainingpapersofStepVweresubmittedforfull-textreading.Tosavethedata,weusedStArttool.Additionally,duringthisstage,weextracteddataaccordingtoanextractionform(seeTable6inAppendixA),whichenabledustorecorddetailsaboutthereviewedpapers.
3. QUALITy ASSeSSMeNT
Thequalityassessmentoftheselectedstudiesisusefultoincreasetheaccuracyofthedataextractionresults.Inaddition,itisimportanttodeterminethevalidityoftheinferencesandinascertaining
Figure 1. Paper selection flow chart
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
73
credibilityandcoherentsynthesisofresults.Table3presentsthequalityassessment(Kitchenhametal.,2007)criteria.Criteria2,4and8onlyreceiveYes(Y)orNo(N)asscore.Additionally,weallowedonlyYes(Y)andPartially(P)scorestothelastcriteriasinceweareinterestedinthepossiblebenefitstoindustry.
TherawdataofthequalityassessmentispresentedinTable72oftheAppendixB,accordingtoassessmentquestiondescribedinTable3.Thescoresof11papersarelessthan50%andtheaveragescoreis7.08seeFigure2.Regardingthescoreofproposedcriteria,Q1,Q2andQ3receivedthehighestaveragesscores(>0.9),whileQ4,Q5,Q8,Q9,Q10andQ11receivedintermediatescoresbetween0.55and0.83.CriteriaQ6andQ7receivedthelowestaveragescore,0.14and0.13respectively.
4. ReSULTS
Thissectionpresents the resultsobtainedby running theSRLmethodpreviouslydescribed.Weprovideadetailedoverviewofgeneralfeaturesofthesestudies:sourcetypes,yearofpublication,andapplicationcontext.ThereviewedpaperswerepublishedbetweenJanuaryof2006andDecemberof2015.Additionally,thedistributionofselectedstudiesoverpublicationsources,includingpublicationname,type,count(i.e.thenumberofselectedstudiesfromeachsource),andthepercentageofselectedstudiesareavailable3.The80selectedstudiesaredistributedin49publicationsources,suggestingthatlinkeddatatoolshavebeenawidespreadconcerninthecommunity.Thestudiesincludedinthisreviewbelongthefollowingcategories:(1)journal;(2)conference;(3)workshop;or(4)bookchapters.Weidentifiedthatmoststudiesbelongtothecategory(2)(38.75%or31studies),followedbycategory(4)(27.5%or22studies),category(3)(17.5%or14studies)andcategory(1)(16.25%or13studies).Wealsoobservedthatmostoftheseselectedworkswerepublishedin2012(22.5%)
Table 3. Quality assessment criteria
# Questions Possible Answers
Q1 Istherearationaleforwhythestudywasundertaken? Y=1,N=0,P=0.5
Q2 Isthepaperbasedonresearch? Y=1,N=0
Q3 Isthereaclearstatementoftheresearchgoals? Y=1,N=0,P=0.5
Q4 Doesthestudyreusesomeframework(software,sourcecode)? Y=1,N=0
Q5 Isthereanadequatedescriptionofthecontextinwhichtheresearchwascarriedout? Y=1,N=0,P=0.5
Q6 Howmanystepsofthepublicationprocessdoesthestudysupport? Number of publication steps
Total of publication steps
Q7 Howmanystepsoftheconsumptionprocessdoesthestudysupport?
Number of consumption steps
Total of consumption steps
Q8 Wasthestudyempiricallyevaluated? Y=1,N=0
Q9 Isthereadiscussionabouttheresultsofthestudy? Y=1,N=0,P=0.5
Q10 Arethelimitationsofthisstudyexplicitlydiscussed? Y=1,N=0,P=0.5
Q11 Doestheresearchalsoaddvaluetotheindustrialcommunity? Y=1,P=0.5
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
74
followedby2010(15%),2013(15%),2014(13.8%),2011(12.7%),2009(6.3%),2008(3.8%),2000(1.3%),2007(1.3%).
ThequestionRQ1identifiesthepurposeofeachsoftwaretool.Weclassifiedthetoolsintotwocategories:publicationandconsumption(seeFigure4).Thesoftwaretoolscanbeclassifiedinoneorbothcategories,accordingtothesupportedsteps.Figure3depictsthat9softwaretools(11.11%)providesupporttooneormorestepsofbothlinkeddataconsumptionandpublicationprocesses.Thesetoolscanbeclassified,underPublicationpointofview,accordingtothefollowingfeatures:RDFpublisher(S01,S19,S25,S37,andS38),RDFsynchronizer(S33),RDFcompactor(S18),LinkedDataconsultinglanguage(S21)andsearchfederatedengine(S27).Table4describesdetailssuchassource-codelanguage,platform,publisher,supportedserializableformats,evaluationmethodandpublicationyear.Figure4depictsthedistributionofthese9studies,inpublicationandconsumptionprocesses.Mostsolutions(77.77%;7studies)arefocusedonprovidingmachineaccess(endpoint,publishingdatasets,etc.)todata(SPARQL,RDFandRDFa).Ontheotherhand,onlyonesoftwaretool(11.11%,1study)isconcernedwiththeconversionofstructureddatatolinkeddataandanotherone(11.11%)isconcernedwiththreesteps:definitionofURIsforLinkedData,announcementofdatasetsandtoestablishsocialcontractofalinkeddatapublisher.
ThequestionRQ2identifiesserializableformatssupportedbysoftwaretools.Figure5showsthe relation between serialization formats and publication, as well as consumption process. WecannotethegrowingsupporttoRDFserializationformats,wherethegreatestamountofsoftwaretoolsrelatedtopublicationprocessfocusesonprovidingmachineaccess(27studies).Ontheotherhand,regardingtotheconsumptionprocess,theconsumptionpatterns(28studies)andalignment(28studies)arethetwomainstepshandled.Itisalsoimportanttopayattentiontothelowsupporttonon-RDFserializationsinStepvii(Convertdata)oftheLDpublicationprocess.InFigure5,itisworthtonotethesupportisconcentratedinthesteps4and5oftheconsumptionprocess.Infact,themostsupportedserializationsareRDF(33.75%,27studies)andSPARQL(12.5%,10studies).Ontheotherhand,themostsupportednon-RDFareXML(8.75%,7studies),CSV(7.5%,6studies)andJSON(5%,4studies).
The question RQ3 enables the identification of evaluated-empirically software tools. Theclassificationoftheseevaluationsisobtainedfromeachpaper.ThenumberofstudiesforeachcategoryisshowninFigure5.Accordingtotheresults,40%ofthestudies(32studies)werenotevaluated,andtheotherstudiesaredistributedbetweenexperiment(36.25%,29studies),usecase(11.25%,9
Figure 2. Overview of the studies. Paper quality assessment distribution.
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
75
Figure 3. Classification of software tools in publication and consumption
Figure 4. Software tools distribution in steps of publication and consumption processes
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
76
studies),proofofconcept(2.5%,2studies),benchmark(1.25%,1study)andsimulation(1.25%,1study).Figure7depictshowtheevaluationsaredistributedalongpublicationandconsumption.Itisworthtonotethat50%(23studies)oftheworksdealingwiththepublicationprocesswerenotevaluated.
5. DISCUSSIoN
Thissectionpresentsadiscussionontheobtainedresults.Weintendtoprovideourperspectiveabouteachresearchquestion,highlightingthefaultsinthecurrentstateofsoftwaretoolsforpublishingandconsumingLinkedData.
Theresultsindicateanunevendistributionofsoftwaretoolsamongthepublicationsteps.Wenotethat63.82%ofthestudieswerefocusedonprovidingmachineaccess,followedbydataconversiontolinkeddataandthedefinitionofURIsforlinkeddata(with34.04%each).Ontheotherhand,wehavedetectedagaponsolutionstopreparestakeholders,supportopendatalicensespecification(2.12%,onestudy),selectadatasetforreuse(10.63%,fivestudies),usestandardsvocabulariesfordescribingobjects(12.76%,sixstudies)andmodelthedatatorepresentdataobjectsandtheirrelationships(24.89%,sevenstudies).Aswecansee,mostoftheproposedsoftwaretoolssupportonlythefinalstepsofthepublicationprocess(44studies)whileonly24studiessupporttheinitialsteps.
Theresultsarepartiallyexplainedbythehigh-levelvisionofthepublicationprocessdetailsprovidedbyW3Cdocuments.AlthoughW3C(2014)pointsouttheimportancetouse/determinedata
Table 4. Publication and consumption software tools
Study Name Developer Year Platform Language Supported Serializable Formats
Evaluation Method
S01 OrganiKproject Industry 2010 WEB PHP RDF Experiment
S18 RDSZ Academy 2014 Notdescribed Python RDF Experiment
S19 RDFextensionforDRUPAL Industry 2000 WEB PHP RDFa,RDFS,OWL
Doesnotpresentempiricalevaluation
S21 FREyA Industry 2012 Notdescribed FREyA SPARQL,RDF Experiment
S25 LinkedLab Academy 2011 WEB Notdescribed
RDF,OWL,XML,JSON,CSV UseCase
S27 KonneXSALT Academy 2008 WEB Java RDF,SPARQL
Doesnotpresentempiricalevaluation
S33 RDFSync Industry 2007 Notdescribed
Notdescribed RDF,SPARQL Experiment
S37 Rhizomer Academy 2011 WEBJava,HTML,JavaScript
SPARQL
Doesnotpresentempiricalevaluation
S38 Bioqueries Academy 2011 WEB PHP RDF,SPARQL,RDFS
Doesnotpresentempiricalevaluation
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
77
license,itlacksdiscussingdetailsonhowtoimplementsuchtaskorwhichvocabularytoadopt.Asnoneanalyzedtoolcoverstheinitialstepsofthepublicationprocess,thereisnotoolabletofulfillthewholepublicationprocess.Basedontheaforementioneddiscussion,wecanpointouttwoquestionsthatcanberaisedforaresearchagenda.
1. HowtodevelopsoftwaretoolstocopewiththeinitialphasesofthepublicationprocessofLinkedData?2. HowtoprovideanintegratedsolutiontosimplifythepublicationofLinkedData?
RegardingtheLDconsumptionprocess,thepresentresultsindicatethatconsumptionsoftwaretools are concentrated in creating consumption patterns to specify what data are reused from acertaindatasource(85%,34studies),followedbymanagingalignments(85%,34studies).Further,thespecificationofusecasesfornewservicesorapplications(2.5%,1study)anddatasourcesanddatasetsevaluation(5%,2studies)aresupportedbyfewstudies,whiletheremainingstepswerenotcoveredbythereviewedstudies.Sincedevelopersandcompaniesusuallyhavespecificobjectivestoreuseadataset,theydeveloptheirownconsumptionpatternsandalignments.Consequently,wecandrawathirdquestionforaresearchagenda:
3. HowtodevelopandsharereusableconsumptionpatternstofacilitatetheconsumptionofLinkedData?
In addition, it is important toprovidegoodURIs4, convertdata to linkeddata, andprovidemachineaccess inpublicationprocess.Considering that thesestepssupportahigherquantityofserializationformats,theresultspointoutadisperseuseofserializableformatsinbothprocesses.Thus,theresultsindicatethatthesupportofserializableformatsimpactsdirectlyonhowinstitutionspublishtheirdata,aswellasconsumersreusethepublisheddataontheWeb.Anotherimportant
Figure 5. Evaluation distribution
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
78
pointisthelowsupportprovidedbythesoftwaretoolstopreviousandsubsequentstepsof“tocreateconsumptionpatternstospecifywhatdataarereusedfromacertaindatasource”and“tomanagealignment,cachingandupdatemechanisms”.
Similarly,tothepublicationprocess,wewerenotabletodetermineanytoolcapableofsupportingthewholeconsumptionprocess.Thus,itbringsusanimportantchallengethatwesummarizeinthefollowingquestiontotheresearchagenda:
4. HowtoprovideanintegratedsolutiontosimplifytheconsumptionofLinkedData?
ThecreationofconsumptionpatternsandthemanagementofalignmentmechanismsarethestepsofLDconsumptionprocessthatmoresupporttheserializationformats.ItisworthtonotethetworectanglespresentedinFigure6,sincetheycontainthehighestnumberofsupportedstepsandserializableformats.Therefore,thelowsupportforserializationformatsimpactsnegativelyinthe
Figure 6. Relation between publication and consumption processes and serializable formats
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
79
consumptionprocess,thusitisnecessarytoproposeapproachestoincreasethenumberofsupportedformats(structuredandRDFserialization).Thisleadsustoafifthquestion:
5. HowtoprovideasolutiontosupportsRDFandnon-RDFserializationformatsinLinkedDataPublicationandConsumptionprocess?
Theempiricalevaluationofanytoolisessentialtoensurequality,aswellastohighlightthetoolfeatures.Nevertheless,almosthalfofevaluatedtools(45%,36studies)donotpresentanykindofevaluation(empiricalornot).Therefore,itisimportanttoquantifytheeffectivenessofeachanalyzedtooltoperformtheintendedtaskproperly,whichbringsustothelastquestion:
6. Howtovalidatetheeffectivenessoftheproposedstudiesaccordingtosupportedsteps?
Inthebigpicture,wecanpointoutanincreasing,albeitbadlydistributed,offerofsoftwaretoolstosupplythepublicationandconsumptionofLinkedData.Inaddition,aconsiderablepartofthesetoolsdoesnotconsiderexistingproductionandconsumptionprocessesduringdevelopment.Thus,itisimportanttoprovideanewsetofsoftwaretoolsthattakeintoaccountthecurrentLinkedDatastateoftheart.
Figure 7. Evaluation by purpose
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
80
Throughoutthisstudy,wepresentedaconsiderablenumberoftoolstoassistthepublisherinbothprocessesofpublishingandconsuminglinkeddata.Thoughtheimportantcontributionspresented,accordingtoBizeretal. (2009) thefollowingquestionsarestillopen:1) thesupportof licenses(Heath&Bizer,2011;Schmachtenberg,Bizer&Paulheim,2014),2)theautomaticevaluationofdataquality(Maα&Qi,2013;Moss,Corsar,&Piper,2012),and3)thedataalignment(Bleiholder&Naumann,2009;Homoceanu,Kalo,&Balke,2014).Besides, it isnoteworthy theneedofanintegrationenvironmenttosimplifytheprocessexecution.
6. THReATS To VALIDITy
AsthreatstothevalidityofthisSLR,themainconstructsinthisreviewarethethreeconcepts“SoftwareTools”,“LinkedData”and“Ontologies”.Forthefirstconcept,weusedtheterms“environment”,“authoring”,“architecture”,“framework”,“software”,“application”and“program”tocollectstudiesrelatedtosoftwaretoolapproaches.Forthesecondconcept,theterms“LinkedData”and“LinkedOpenData”wereusedtocoverthepotentiallyrelevantstudiesonLinkedDatafromthedatabasesearch.Forthethirdconcept,theterms“ontology”anditsplural“ontologies”wereusedtocollectstudiesrelatedtoontologyapproaches.However,wedidnotusecontrolpapersduringthesearchondigitallibraries,whichmaybeconsideredasthemainthreattovalidity.Additionally,wedidnotperformacomplementarymanualsearch.
7. CoNCLUSIoN
Inthiswork,weconductedaSLRtoinvestigatethesupportofsoftwaretoolstopublishandconsumeLinkedData.Ourgoalwastoimprovetheunderstandingofhowthesesoftwaretoolssupporteachstepofbothprocesses.
Eightystudieswerefullyinvestigated,ofwhichnineprovidesupportinoneormorestepsoftheconsumptionandpublicationprocesses.Inregardstothepublicationprocess,itispossibletoevidencethelowsupportintheinitialsteps.Moreover,intheconsumptionprocess,itispossibletoevidencethesoftwaretoolsfocusthesupportofthedevelopmentofconsumptionpatterns,aswellasdataalignment.
Aspreviouslydiscussed,weraisedasetofquestions,whichinvolvesthestepsinprocesses,software tools,serializationformatsandtherelationsamongthem.Finally,wehighlightedsomeimportantquestionsthatarerevisedandsummarizedinTable5inordertosuggestaresearchagendaonlinkeddata.
Basedon(Kitchenhametal.,2007)wecanhighlightsomecontributionsprovidedbythisstudy.Suchcontributionsinclude:
• Theidentificationofgapsincurrentresearchinordertosuggestareasforfurtherinvestigation;• Theprovisionofaframework/backgroundinordertoproperlypositionnewresearchactivities;• Thesummarizationofexistingevidence,benefitsandlimitations.
The result presented in thisSystematicLiteratureReviewcanbeveryuseful to linkeddatacommunity,sinceitgathersevidencesfromprimarystudiesincludedinthereview,formingabodyofknowledgeregardingtheuseofsoftwaretoolstopublishorconsumelinkeddata.Wecanoutlinethecontributionofthisworkasfollows:
• ResearchagendabasedonSRLobservationsandaccordingtogapsinbothprocesses.Besides,wepointedouttheneedofanintegrationenvironmenttosimplifytheprocessexecution.
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
81
Asfuturework,weintendtofurtherinvestigatethesoftwaretoolintegration.Moreover,weintendtoextendthissystematicreviewaimingtoclarifythecommunicationinterfaceamongsoftwaretoolsbetweenstepsineachprocess.
Table 5. Questions and motivation
Motivation Question
Theinitialstepofthepublicationprocessisbadlyexploredbytheavailabletools.
HowtodevelopsoftwaretoolstocopewiththeinitialphasesofthepublicationprocessofLinkedData?
Nonesolutionsupportsthewholeprocessofpublicationofconnecteddata.
HowtoprovideanintegratedsolutiontosimplifythepublicationofLinkedData?
Thereisnotastandardandin-depthapproachtotheconsumptionprocessofconnecteddata.
HowtodevelopandsharereusableconsumptionpatternstofacilitatetheconsumptionofLinkedData?
Nonesolutioncoversthewholeprocessofdataconsumption.
HowtoprovideanintegratedsolutiontosimplifytheconsumptionofLinkedData?
Thereisneedtoincreasethenumberofsupportedserializationformatsforbothpublicationandconsumptionprocesses.
HowtoprovideasolutiontosupportsRDFandnon-RDFserializationformatsinLinkedDataPublicationandConsumptionprocess?
Lownumberofsoftwaretoolswasevaluated.Additionally,theseevaluationswarenotfocusedonsupportpublicationorconsumptionprocess.
Howtovalidatetheeffectivenessoftheproposedstudiesaccordingtosupportedsteps?
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
82
ReFeReNCeS
Auer,S.,Bühmann,L.,Dirschl,C.,Erling,O.,Hausenblas,M.,Isele,R.,&Williams,H.(2012).Managingthelife-cycleoflinkeddatawiththeLOD2stack.InThe Semantic Web–ISWC 2012(pp.1–16).SpringerBerlinHeidelberg.doi:10.1007/978-3-642-35173-0_1
Barnes,M.,&Martin,R.(2002).Businessdatalinking:Anintroduction.Economic Trends,581,34–41.
Bauer,F.,&Kaltenböck,M.(2011).Linkedopendata:Theessentials.Editionmono/monochrom,Vienna.
Bizer,C.,Cyganiak,R.,&Heath,T.(2007).HowtopublishLinkedDataontheWeb.RetrievedSeptember18,2016,fromhttp://wifo5-03.informatik.uni-mannheim.de/bizer/HowtoPublishLinkedData.htm
Bizer,C.,Heath,T.,&Berners-Lee,T.(2009).LinkedData-TheStorySoFar.International Journal on Semantic Web and Information Systems,5(3),1–22.doi:10.4018/jswis.2009081901
Bleiholder, J., & Naumann, F. (2009). Data fusion. ACM Computing Surveys, 41(1), 1–41.doi:10.1145/1456650.1456651
Corlosquet,S.,Delbru,R.,Clark,T.,Polleres,A.,&Decker,S.(2009).ProduceandConsumeLinkedDatawithDrupal! InProceedings of the International Semantic Web Conference (pp.763-778).SpringerBerlinHeidelberg.doi:10.1007/978-3-642-04930-9_48
Dermeval,D.,Vilela,J.,Bittencourt,I.I.,Castro,J.,Isotani,S.,&Brito,P.(2014,September).Asystematicreviewontheuseofontologiesinrequirementsengineering.InProceedings of the 2014 Brazilian Symposium on Software Engineering (SBES)(pp.1-10).IEEE.doi:10.1109/SBES.2014.13
Dietze,S.,Yu,H.Q.,Giordano,D.,Kaldoudi,E.,Dovrolis,N.,&Taibi,D.(2012,March).LinkedEducation:interlinkingeducationalResourcesandtheWebofData.InProceedings of the 27th annual ACM symposium on applied computing(pp.366-371).ACM.doi:10.1145/2245276.2245347
Ding,L.,Lebo,T.,Erickson,J.S.,DiFranzo,D.,Williams,G.T.,Li,X.,&Hendler,J.A.(2011).TWCLOGD:Aportalforlinkedopengovernmentdataecosystems.Web Semantics: Science, Services, and Agents on the World Wide Web,9(3),325–333.doi:10.1016/j.websem.2011.06.002
Heath,T.,&Bizer,C.(2011).Linkeddata:Evolvingthewebintoaglobaldataspace.Synthesis lectures on the semantic web: theory and technology,1(1),1-136.
Hernandes,E.,Zamboni,A.,Fabbri,S.,&Thommazo,A.D.(2012).UsingGQMandTAMtoevaluateStArt-atoolthatsupportsSystematicReview.CLEI Electronic Journal,15(1),3–3.
Homoceanu,S.,Kalo,J.C.,&Balke,W.T.(2014,June).PuttingInstanceMatchingtotheTest:IsInstanceMatchingReadyforReliableDataLinking?InProceedings of theInternational Symposium on Methodologies for Intelligent Systems(pp.274-284).SpringerInternationalPublishingdoi:10.1007/978-3-319-08326-1_28
Hyland,B.,&Wood,D.(2011).Thejoyofdata-acookbookforpublishinglinkedgovernmentdataontheweb.InLinking government data(pp.3–26).SpringerNewYork.doi:10.1007/978-1-4614-1767-5_1
Kitchenham,B.,&Charters,S. (2007).Guidelinesforperformingsystematic literaturereviews insoftwareengineering.InTechnical report, Ver. 2.3.EBSETechnicalReport.EBSE.
LaPES.(n.d.).Start-stateoftheartthroughsystematicreviewtool.RetrievedSeptember04,2016,fromhttp://lapes.dc.ufscar.br/tools/start_tool
Moss,L.,Corsar,D.,&Piper,I.(2012,June).Alinkeddataapproachtoassessingmedicaldata.InProceedings of the 2012 25th International Symposium on Computer-Based Medical Systems (CBMS) (pp. 1-4). IEEE.doi:10.1109/CBMS.2012.6266391
Schmachtenberg,M.,Bizer,C.,&Paulheim,H.(2014,October).Adoptionofthelinkeddatabestpracticesindifferenttopicaldomains.InProceedings of theInternational Semantic Web Conference(pp.245-260).SpringerInternationalPublishing.doi:10.1007/978-3-319-11964-9_16
Shadbolt,N.,OHara,K.,Berners-Lee,T.,Gibbins,N.,Glaser,H.,Hall,W.,&schraefel,.(2012).Linkedopengovernmentdata:Lessonsfromdata.gov.uk.IEEE Intelligent Systems,27(3),16–24.doi:10.1109/MIS.2012.23
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
83
Armando Barbosa is a Master’s student in Informatics graduate program, Computing Institute at the Federal University Alagoas (Brazil) and a member of the Center of Excellence for Social Technologies. Armando’s main research line is Semantic Web, Linked Data.
Ig Ibert Bittencourt is Professor at the Federal University of Alagoas (Brazil) and CoDirector of the Center of Excellence for Social Technologies. He received his PhD in 2009 from the Federal University of Campina Grande, Brazil and his Post-Doctoral degree in 2013 at the University of Campinas (UNICAMP). He is also the vice-president of the Special Committee of Computers and Education from Brazilian Computer Society and W3C Advisory Committee Representative. Ig Bittencourt also founded a Company in Intelligent Tutoring Systems awarded as the most innovative Brazilian company in 2014 (USP Innovation Prize). He believes in innovative social entrepreneurship as a model for promoting a sustainable economic and social development. His main interests are research, development and innovation in Computers and Education, Ontologies and Social Entrepreneurship.
Sean W.M. Siqueira is Professor at the Department of Applied Informatics, Federal University of the State of Rio de Janeiro (UNIRIO), Brazil, where he teaches courses in Information Systems, Databases, Web Science and Social & Semantic Web. He holds an MSc (1999) and a PhD (2005) in Computer Science, both from the Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Brazil. His research interests include knowledge representation and management; collaborative systems; social web; semantic web; ontologies; information integration; semantic models; user models; and learning & education. He has experience in the Computer Science area, with focus on Web Science, Information Systems and Technology Enhanced Learning. He has participated in some international research projects and has written more than 130 papers for conferences, journals, and book chapters. He was the coordinator of the Graduate Program in Information Systems at UNIRIO from July 2012 to September 2014, the coordinator of the Program Committees of the Brazilian Symposium on Information Systems (SBSI 2015), the Brazilian Symposium on Computers and Education (SBIE 2012 and SBIE 2014), member of the steering committee of the Brazilian Congress of Computers and Education (CBIE 2015), co-editor-in-chief of the iSYS: Brazilian Journal of Information Systems as well as the Brazilian Journal on Computers in Education, and member of the special committee on Computers and Education (CEIE) and coordinator of the special committee on Information Systems (CESI), both from the Brazilian Computer Society.
Rafael de Amorim Silva received his BS degree in Information System from the Faculty of Alagoas (2005), Brazil. He received his MS degree in Computer Science from the Federal University of Pernambuco (2008), Brazil, and his PhD degree in Electronics Engineering from the Aeronautics Institute of Technology, Brazil. He is currently professor of the Federal University of Alagoas, Brazil. His main scientific interests include Computer Networks, Informatics in Education, and Artificial Intelligence.
Ivo Calado holds a BSc (2008) and MSc (2010) in Computer Science at Federal University of Alagoas and Federal University of Campina Grande, respectively. He also holds a PhD in Electrical Engineering at Federal University of Campina Grande, Brazil (2015). Currently, Ivo Calado is a Professor at Federal Institute of Education, Science and Technology of Alagoas, Brazil. His main research interests include Computer Networks, Internet of Things and Semantic Web.
Villazón-Terrazas, B., Vilches-Blázquez, L. M., Corcho, O., & Gómez-Pérez, A. (2011). Methodologicalguidelinesforpublishinggovernmentlinkeddata.InLinking government data(pp.27–49).SpringerNewYork.doi:10.1007/978-1-4614-1767-5_2
WorldWideWebConsortium.(2014).Bestpracticesforpublishinglinkeddata.RetrievedSeptember04,2016,fromhttps://www.w3.org/TR/ld-bp/
eNDNoTeS
1 https://www.w3.org/2001/sw/wiki/Category:Tool2 http://bit.ly/2c7UlHc3 http://bit.ly/2cB9CzE4 http://www.w3.org/TR/ld-bp/#HTTP-URIS
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
84
APPeNDIX A
extraction Form
Table 6. Extraction form
# Study Data Description Relevant RQ
1 StudyIdentifier Uniqueidforthestudy Studyoverview
2 Dateofdataextraction,Authors,Year,Title,Country Studyoverview
3 ArticleSource Studyoverview
4 TypeofArticle Journal,Conference,workshop,bookchapter Studyoverview
5 Softwaretoolname Whatarethesoftwaretoolname? Studyoverview
6 Version Whatarethesoftwaretoolversion? Studyoverview
7 LastUpdate Whenwasthelastupdateofthesoftwaretool? Studyoverview
8 License Whatisthelicense? Studyoverview
9 Platform Web,Mobile,Desktop Studyoverview
10 Language Whatare/istheusedlanguagetoimplementthesoftwaretool? Studyoverview
11 ApplicationContext Industrial,Academic Studyoverview
12 Purpose Publication,Consumption RQ1
13 Stepsfollowed Whatarethestepsfollowedofthepurpose? RQ1
14 Serializableformats Whataretheserializableformatssupportedbythesoftwaretool? RQ2
15 ResearchMethod Controlledexperiment,Casestudy,Usecase,Conceptualproof,Benchmark,Simulation,Notapplicable RQ3
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
85
APPeNDIX BQuality Assessment
Table 7. Quality assessment
ID References Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Total Score Quality
S01 (Aastrand,Celebi,&Sauermann,2010) 1 1 1 1 1 0,4 0,3 1 1 0,5 1 9,2 84%
S02 (An,Kim,Lee,&Lee,2013) 1 1 1 1 0,5 0 0,3 1 1 1 1 8,8 80%
S03
(Dimou,DeVocht,VanGrootel,VanCampe,Latour,Mannens&VandeWalle,2014)
1 1 1 1 1 0 0,1 1 1 0 0,5 7,6 69%
S04
(Lehmann,Furche,Grasso,Ngomo,Schallhart,Sellers,&Auer,2012)
0 0 1 0 0 0,1 0 0 1 1 1 4,1 37%
S05 (Aprosio,Giuliano,&Lavelli.2013) 1 1 1 0 0,5 0,3 0 1 1 1 0,5 7,3 66%
S06(Araujo,Houben,Schwabe,&Hidders,2010)
1 1 1 1 0 0 0,4 0 1 0 0,5 5,9 54%
S07(Assaf,Louw,Senart,Follenfant,Troncy,&Trastou,2012)
1 1 1 1 1 0,2 0 1 1 1 1 9,2 84%
S08(Auer,Dietzold,Lehmann,Hellmann,&Aumuelle,2009)
1 1 1 0 1 0,4 0 1 1 1 1 8,4 76%
S09 (Augenstein,Padó,&Rudolph,2012) 1 1 1 1 0,5 0,5 0 1 1 1 1 9 82%
S10 (Lehmann&Bühmann,2011) 1 1 1 1 0,5 0,1 0 1 1 1 0,5 8,1 74%
S11 (Bottoni&Cerian,2014) 1 1 1 0 0,5 0 0,4 0 0 0 0,5 4,4 40%
S12 (Chang,Xu,Zhou,Shao,Li,&Yan,2013) 1 1 1 1 0,5 0,4 0 0 0 1 0,5 6,4 58%
S13
(Chen,Ding,Wang,Wild,Dong,Sun,&Sankaranarayanan,2010)
1 1 1 1 0,5 0,4 0 1 1 1 0,5 8,4 76%
S14 (Becker&Bizer,2009) 0,5 1 1 0 0,5 0 0,3 0 0 1 0,5 4,8 44%
S15
(Bizer,Lehmann,Kobilarov,Auer,Becker,Cyganiak,&Hellmann,2009)
1 1 1 1 0,5 0,4 0 0 1 1 0,5 7,4 67%
S16 (Ciravegna,Gentile,&Zhang,2012) 1 1 1 0 0,5 0 0,3 0 0 0 0,5 4,3 39%
S17
(Coletta,Castanier,Valduriez,Frisch,Ngo,&Bellahsene,2012)
1 1 1 1 0,5 0,2 0 0 1 0 0,5 6,2 56%
continued on following page
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
86
ID References Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Total Score Quality
S18
(Fernández,Arias,Sánchez,Fuentes-Lorenzo,&Corcho,2014)
1 1 1 1 0,5 0,1 0 1 1 1 1 8,6 78%
S19 (Corlosquetetal.,2009) 1 1 1 1 1 0,4 0 1 1 0 0,5 7,9 72%
S20 (Tao,Song,Sharma,&Chute,2013) 1 1 1 1 0,5 0,1 0 1 1 1 0,5 8,1 74%
S21(Agatonovic,Cunningham,&Damljanovic,2012)
1 0 1 1 0,5 0,1 0 1 1 0 0,5 6,1 55%
S22(Curry,O’Donnell,Corry,Hasan,Keane,&O’Riain,2013)
1 1 1 0 1 0,3 0 0 1 0 1 6,3 57%
S23(DiNoia,Mirizzi,Ostuni,Romito,&Zanker,2012)
0,5 0 0 0 0,5 0 0,3 1 1 0 0,5 3,8 34%
S24(DeVocht,Softic,Mannens,Ebner,&VandeWalle,2014)
1 1 1 0 1 0 0,3 1 1 1 0,5 7,8 71%
S25 (Darari&Manurung,2011) 0,5 1 1 0 0,5 0,5 0 0 1 1 0,5 6 55%
S26 (Dastgheib,Kochut,&Mesbah,2013) 1 1 1 1 0,5 0 0,1 0 1 1 0,5 7,1 65%
S27(Groza,Handschuh,Möller,&Decker,2008)
1 1 1 1 1 0,2 0 0 1 1 0,5 7,7 70%
S28 (Delbru,2009) 1 1 1 0 1 0 0,3 1 1 0 0,5 6,8 62%
S29 (Auer,Dietzold,&Doehring,2010) 1 1 1 1 0,5 0,1 0,3 1 1 1 1 8,9 81%
S30(Paret,VanWoensel,Casteleyn,Signer,&DeTroyer,2011)
1 1 1 0 0,5 0 0,3 1 1 1 0,5 7,3 66%
S31 (Elze,Hesse,&Martin,2011) 1 1 1 1 1 0 0,3 0 1 1 0,5 7,8 71%
S32(Emaldi,Lázaro,Laiseca,&López-de-Ipiña,2012)
0,5 1 0,5 1 0,5 0 0,3 1 1 1 0,5 7,3 66%
S33
(Tummarello,Morbidoni,Bachmann-Gmür,&Erling,2007)
1 1 1 0 0,5 0,3 0 1 1 1 1 7,8 71%
S34 (Baritakis,Fafalios,&Tzitzikas,2014) 1 1 1 1 0,5 0 0,3 1 1 0 0,5 7,3 66%
S35 (Farouk&Ishizuka,2012) 1 1 1 0 0,5 0,2 0 1 1 1 0,5 7,2 65%
S36(Fernández-Tobías,Cantador,Kaminskas,&Ricci,2011)
0,5 1 0,5 0 0,5 0 0,3 0 1 1 0,5 5,3 48%
S37(García,Brunetti,López-Muzás,Gimeno,&Gil,2011)
1 1 1 1 0,5 0,1 0 0 1 1 0,5 7,1 65%
Table 7. Continued
continued on following page
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
87
ID References Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Total Score Quality
S38
(Aldana-Montes,García-Godoy,&Navas-Delgado,2011)
1 1 1 1 1 0,4 0 0 1 1 0,5 7,9 72%
S39 (Graves,2010) 1 1 1 1 0,5 0,4 0 0 1 1 0,5 7,4 67%
S40(Wang,Liu,Penin,Fu,Zhang,Tran,&Pan,2009)
1 1 1 1 0,5 0 0,3 1 1 1 0,5 8,3 75%
S41 (Chen,Hsueh,&Huang,2011) 1 1 1 0 1 0,1 0 0 0 0 0,5 4,6 42%
S42(Fisteus,García,Fernández,&Fuentes-Lorenzo,2014)
1 1 1 1 0,5 0,1 0 1 1 1 0,5 8,1 74%
S43 (Han,Finin,Parr,Sachs,&Joshi,2008) 0,5 1 1 0 0,5 0,4 0 1 0 1 0,5 5,9 54%
S44 (Bauer,Kaltenböck,&Recheis,2011) 1 1 1 0 0,5 0,4 0 0 0 0 0,5 4,4 40%
S45 (Auer,Hladky,&Khalili,2012) 1 1 1 1 0,5 0,1 0 1 1 1 0,5 8,1 74%
S46 (Auer&Khalili,2013) 1 1 1 1 0,5 0,1 0 1 1 1 0,5 8,1 74%
S47 (Consens&Khatchadourian,2010) 1 1 1 0 0,5 0 0,3 1 1 0 0,5 6,3 57%
S48 (Grollios,Mitkas,&Vavliakis,2013) 1 1 1 0 0,5 0 0,1 1 1 1 0,5 7,1 65%
S49 (Konstantinou,Kouis,&Mitrou,2014) 1 1 1 1 0,5 0,4 0 1 1 1 0,5 8,4 76%
S50(Hernandez,Lappalainen,&Sicilia,2013)
1 1 1 0 1 0 0,7 0 0 1 0,5 6,2 56%
S51
(Lehmann,Furche,Grasso,Ngomo,Schallhart,Sellers,&Auer,2012)
1 1 1 0 0,5 0 0,1 1 1 0 0,5 6,1 56%
S52 (Lowe,2009) 1 1 1 1 1 0,2 0 1 1 0 0,5 7,7 70%
S53(Lynden,Kojima,Matono,&Tanimura,2011)
1 1 1 1 0,5 0,2 0 1 1 1 0,5 8,2 75%
S54 (Machado&ParentedeOliveira,2011) 1 1 1 0 0,5 0,4 0 0 1 1 1 6,9 63%
S55 (Narasimha,Kappara,Ichise,&Vyas,2011) 1 1 1 1 0,5 0 0,3 1 1 1 0,5 8,3 75%
S56 (Curé,2014) 1 0 1 1 1 0 0,1 0 1 1 0,5 6,6 60%
S57(Ostuni,DiNoia,Mirizzi,Romito,&DiSciascio,2012)
0,5 1 1 0 0,5 0 0,3 0 1 1 0,5 5,8 53%
S58(Otero-Garcıa,Vidal,Lama,Bugarın,&Domenech)
1 1 1 0 1 0 0,3 0 1 0 0,5 5,8 53%
S59 (Merialdo&Papotti,2012) 1 1 1 0 0,5 0,4 0 1 1 1 0,5 7,4 67%
Table 7. Continued
continued on following page
International Journal on Semantic Web and Information SystemsVolume 13 • Issue 4 • October-December 2017
88
ID References Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Total Score Quality
S60 (Duval&ParraKlerkx,2010) 1 1 1 1 1 0 0,1 1 1 1 0,5 8,6 79%
S61 (Passant,2010) 0,5 0 1 0 0,5 0 0,3 1 1 1 0,5 5,8 53%
S62 (Brambilla,Ceri,&Quarteroni,2013) 0,5 1 1 1 0,5 0,2 0 1 1 0 0,5 6,7 61%
S63
(Radzimski,Sánchez-Cervantes,Rodríguez-González,Gómez-Berbís,&García-Crespo,2012)
1 1 1 1 0,5 0,2 0 0 0 0 0,5 5,2 47%
S64 (Seguran,Senart,&Trastour,2012) 1 1 1 0 1 0,1 0 0 0 0 1 5,1 46%
S65 (Shakya,Takeda,&Wuwongse,2009) 1 1 1 1 0,5 0,1 0 0 0 1 0,5 6,1 55%
S66 (Chapman,Skarka,&Solanki,2012) 1 1 1 0 1 0,1 0 0 1 1 0,5 6,6 60%
S67(Solomou,Kalou,Koutsomitropoulos,&Papatheodorou,2011)
1 1 1 1 0,5 0 0,3 1 1 1 1 8,8 80%
S68(Sorrentino,Bergamaschi,Fusari,&Beneventano,2013)
1 1 1 1 1 0,1 0 0 0 1 1 7,1 65%
S69 (Stadtmüller,Speiser,Harth,&Studer,2013) 1 1 1 0 0,5 0 0,3 1 1 1 1 7,8 71%
S70 (Stolz&Hepp,2013) 1 1 1 1 0,5 0 0,3 1 1 1 1 8,8 80%
S71 (Norheim,Roman,&Stuhr,2011) 1 1 1 1 0,5 0,1 0 1 1 1 1 8,6 78%
S72 (d’Aquin,Motta,&Tiddi,2014) 1 1 1 0 0,5 0 0,1 1 1 1 0,5 7,1 65%
S73 (Tramp,Heino,Auer,&Frischmuth,2010) 1 1 1 0 0,5 0,1 0 1 1 1 0,5 7,1 65%
S74
(Tummarello,Cyganiak,Catasta,Danielczyk,Delbru,&Decker,2010)
1 0 1 0 0,5 0 0,3 0 0 1 0,5 4,3 39%
S75 (Veres,2012) 0,5 1 1 0 0,5 0 0,1 1 1 1 0,5 6,6 60%
S76(Montoya,Ibáñez,Skaf-Molli,Molli,&Vidal,2014)
1 1 1 0 0,5 0 0,3 1 1 1 0,5 7,3 66%
S77(Gómez-Pérez,Vila-Suero,&Villazón-Terrazas,2012)
1 1 1 1 0,5 0,4 0 1 1 1 1 8,9 81%
S78 (Zhang,Song,Li,&Liu,2014) 1 1 1 0 0,5 0,5 0 1 1 1 0,5 7,5 68%
S79
(Kaoudi,Koubarakis,Kyzirakos,Miliaraki,Magiridou,&Papadakis-Pesaresi,2010)
1 1 1 1 0,5 0 1 1 1 1 0,5 9 82%
S80 (Čerāns&Būmans,2010) 1 1 1 1 0,5 0,1 0 1 1 1 0,5 8,1 74%
Average 0,92 0,92 0,97 0,55 0,61 0,14 0,13 0,62 0,83 0,74 0,61 7,08 64%
Table 7. Continued